Histone deacetylase inhibitors (HDACi) represent a relatively new and interesting class of anticancer agent which, as their nomenclature would suggest, inhibit the enzymes that deacetylate histone proteins.1, 2 Histones play an important role in maintaining chromatin structure, and posttranslational modifications to these molecules serve to regulate chromatin density. As a result of histone deacetylase (HDAC) inhibition, histones become hyperacetylated and DNA is maintained in a relatively open conformation that is conducive to interaction with transcription factors. Consistent with this scenario, HDACi have been shown to alter the transcription of a number of genes. In addition, these compounds have been shown to mediate tumor cell differentiation, growth-inhibition and death, and a number of such HDACi are in clinical trials with one (Zolinza™) recently receiving FDA approval for the treatment of cutaneous T-cell lymphoma.
In contrast to the relatively detailed understanding that has been attained regarding the role that histones and their modifications play in maintaining chromatin structure, the mechanism(s) by which HDACi mediate anticancer effects are poorly understood. The discovery that HDACs are inappropriately overexpressed or associated with oncogenic transcription factors in some cancers supports the theory that HDACi may in some cases exhibit anticancer activity by counteracting tumor-associated increases in HDAC activities or by depriving cancer cells of HDAC activities which they have oncogenically subverted. However, many HDACi exhibit potent growth-inhibitory/cytotoxic activity across a wide variety of cancer cell types in a nearly ubiquitous fashion in preclinical studies, thus raising the possibility that aberrant tumor-associated HDAC activity is not a prerequisite for HDACi responsiveness.
While histones clearly represent a predominant class of proteins which exhibit HDACi-mediated hyperacetylation, a number of nonhistone proteins have also been shown to be hyperacetylated and functionally influenced as a consequence of HDACi. Thus, HDACi mediate a variety of transcriptional and nontranscriptional alterations in cancer cells. The ability of HDACi to affect the activity of a variety of disparate proteins and pathways in tumor cells makes them attractive anticancer agents and also makes the understanding of their mechanism of action a daunting task.
We have been studying a novel hydroxamate, belinostat (previously called PXD101), which inhibits HDAC activity in cancer cell lysates with nM IC50 potency and induces the hyperacetylation of proteins such as histones and tubulin in tumor cells.3, 4 This compound, which is considered a pan-HDACi on account of its inhibitory activity on a number of purified recombinant HDACs, was previously shown to inhibit tumor cell growth in vitro and in animal models both as monotherapy as well as in combination with other anticancer agents3–6 and it is currently being evaluated as an anticancer agent in clinical trials. To support the clinical development of belinostat, we performed preclinical studies to evaluate the activity of this compound in prostate cancer.
With an estimated 234,469 new cases and 27,350 deaths in 2006 in the United States alone, prostate cancer ranks as the third leading cause of cancer-related deaths among American men.7 Although significant advances have been made which facilitate the detection of this type of cancer, the treatment of metastatic prostate cancer remains problematic and new therapies are sorely needed. Herein, we report our preclinical findings which support the clinical evaluation of belinostat for the treatment of prostate cancer.
Material and methods
Prostate cancer cell lines were obtained from the American Type Culture Collection (Manassas, VA) and cultured in RPMI-1640 supplemented with 10% FBS. Normal primary prostate epithelial cells were obtained from Cambrex Corp. (East Rutherford, NJ) and cultured in supplemented PREGM media as recommended by the supplier.
In vitro growth-inhibition/cytotoxicity assays on cancer cell lines
Cells were seeded in white 96-well plates (Corning, Acton, MA) at 3 × 103 cells/well and allowed to attach overnight. The following day, cells were treated with 9 different doses of belinostat and then assessed for viability/growth 3 days later using the CellTiter-Glo assay (Promega, Madison, WI). Belinostat has been described3 and for in vitro use was prepared as a 100 mM stock in dimethylsulfolxide (DMSO) and diluted in growth media. Results were recorded on a luminometer from triplicate wells. Following subtraction of background luminescent units, the percent growth inhibition was calculated by dividing the average luminescent units from drug-treated samples by that determined from control wells incubated in the absence of drug. IC50 values were calculated using PRISM software (GraphPad Software, San Diego, CA) with sigmoidal dose-response curve fitting.
Trypan blue dye exclusion viability assay
Cells were seeded in 6-well tissue culture dishes at a subconfluent density and allowed to attach overnight. The following day, cells were treated with belinostat in growth media. In some experiments, belinostat-containing media was washed off at various times and replaced with fresh drug-free growth media. For viability counting, adherent cells from triplicate wells were trypsinized, mixed with trypan blue dye and counted on a haemocytometer. Only viable cells which excluded trypan blue were counted.
Cell cycle analysis
Cells were seeded in 100-mm tissue culture dishes at a subconfluent density and allowed to attach overnight. The following day, cells were treated with belinostat in growth media for 24–48 hr and then processed for cell cycle analysis. For processing, cells were collected, resuspended in 70% ethanol and incubated at −20°C for a minimum of 24 hr (fixation step). After this incubation, cells were centrifuged, resuspended in PBS/0.1% Tween-20/2% FBS and incubated at room temperature for 15–30 min. Following this incubation, cells were centrifuged, resuspended in propidium iodide/RNase staining solution (BD PharMingen, San Diego, CA) and incubated at room temperature in the dark for 15–30 min. Flow cytometry was performed using a FACS Caliber machine (BD Immunocytometry Systems, San Jose, CA) with CellQuest DNA analysis acquisition software and ModFit LT cell cycle analysis software.
To initiate orthotopic xenografts, 5 × 104 exponentially growing PC-3 cells were implanted into the posterior lobe of the prostate of 4- to 5-week-old male athymic nude mice (nu/nu mice; Charles River Laboratories, Wilmington, MA). Ten days after implantation, 5 mice were sacrificed to confirm tumor establishment and the remaining mice were randomized into 4 groups (n = 17 animals/group) and drug treatments were initiated. All treatments were administered into the intraperitoneal cavity (i.p.). The drug stock for in vivo use consisted of belinostat (50 mg/mL) dissolved in water (pH 9.4) containing L-arginine (100 mg/mL). Dosing solutions were prepared by diluting the belinostat stock with PBS. One group of animals (control) was administered vehicle every 12 hr [i.e., twice a day (bid)]. Other groups were administered belinostat either at 20 or 40/mg/kg/dose, also bid. The final group was administered belinostat at 40 mg/kg/dose every 8 hr [i.e., 3 times a day (tid)]. Beginning 10 days after tumor cell implantation, belinostat was administered on days 10–14, 17–21 and 24–28. At the termination of the study, mice were euthanized under CO2 inhalation as per IACUC guidelines and the tumors were excised from the prostates and weighed. Lungs were also removed, fixed with 10% formaldehyde, stained with Bouins solution and examined under a dissecting microscope for evidence of gross metastatic foci. These experiments were performed by KARD Scientific (Wilmington, MA) following the recommendations outlined in the Guide for Care and Use of Laboratory Animals with respect to restraint, husbandry, surgical procedures, feed and fluid regulation and veterinary care.
PC-3 cells were seeded in tissue culture dishes at a subconfluent density and allowed to attach overnight. The following day, belinostat diluted in growth media was added and the cells were returned to the incubator. After 24 hr, cells were trypsinized, washed twice with serum-free RPMI and resuspended in this media at 2 × 105 cells/mL. Cells were then seeded at 0.2 mL/well onto Transwell filters (Falcon™ FluoroBlok Multiwell 8-μm pore size 24-well insert system; BD Biosciences, Bedford, MA) that had been previously coated for 1 hr at 37°C with 10 μg/mL type I rat tail collagen (BD Biosciences). The bottom chambers were filled with 0.9 mL RPMI/1% FBS. Cells were allowed to migrate for 4 hr at 37°C, after which time the media was removed and the filters were transferred to a 24-well plate filled with 0.6 mL HBSS (Invitrogen) containing 5 μM Calcein-AM (Invitrogen) per well. Cells were incubated for 1 hr at 37°C to absorb and process the Calcein-AM. The fluorescent intensity was measured using a PE Biosystems CytoFluor 4000 plate reader at 460/525 nm excitation-emission wavelengths.
Cells grown in monolayer culture were treated as indicated and then harvested in 1 × Tris-glycine SDS sample buffer (Invitrogen, Carlsbad, CA) supplemented with 10 mM dithiothreitol (Sigma, St. Louis, MO). Cell lysates were boiled for 10 min, resolved on 4–20% gradient polyacrylamide gels (Invitrogen) and transferred to nitrocellulose filters (Invitrogen). Immunoblotting was performed using standard procedures and the following primary antibodies: anti-TIMP-1 (R & D Systems, Minneapolis MN), anti-p53 (mix of antibodies from Santa Cruz Biotechnologies, Santa Cruz, CA and EMD Chemicals, San Diego, CA), anti-ERG (Santa Cruz Biotechnology), anti-p21 (BD Pharmingen) and anti-actin (Sigma). Following incubation with the appropriate horseradish peroxidase-conjugated secondary antibodies, enhanced chemiluminescence (GE Healthcare, Chalfont St. Giles, United Kingdom) was used for detection. Actinomycin D, cyclohexamide and emetine were obtained from Sigma and used at final concentrations of 100 ng/mL, 1 μg/mL and 1 μg/mL, respectively. For siRNA experiments, cells were transfected with 100 nM of siGENOME SMART pool reagents (Dharmacon, Chicago, IL) using Oligofectamine tansfection reagent (Invitrogen) following the instructions of the manufacturer.
Growth-inhibitory activity of belinostat on prostate cancer cells
The growth-inhibitory activity of belinostat on 4 human prostate cancer cell lines was determined by exposing cells grown in monolayer culture to 9 drug concentrations for 72 hr followed by examination of cell viability/growth via the CellTiter-Glo assay, which measures ATP (Fig. 1). This analysis yielded the following IC50 values: 22Rv1 (0.18 μM); PC-3 (0.63 μM); LNCaP (0.76 μM); DU145 (0.95 μM). Belinostat (NSC no. 726630) also exhibited sub-μM potency on 2 prostate cancer cell lines (PC-3; DU145) examined by the NCI in a cell viability/growth assay employing sulforhodamine B (NCI data obtained with belinostat on prostate and other cancers cell lines is available at http://dtp.nci.nih.gov).
Cytotoxic activity of belinostat on prostate cancer cells
The effectiveness of an anticancer drug may be related to its ability to promote the death of cancer cells. Since the aforementioned CellTiter-Glo assay does not adequately distinguish growth-inhibition from cell death, we examined the activity of belinostat on prostate cancer cells using trypan blue dye which stains only dead or dying cells. Two prostate cancer cell lines (22Rv1 and PC-3) were exposed to belinostat and viable cell counts were measured each day for 3 consecutive days (Fig. 2). The results of this analysis demonstrated that 22Rv1 cells were completely growth-inhibited by belinostat (1 μM), and that higher drug concentrations (4, 16 μM) induced the death of these cells, as indicated by a decrease in the number of viable cells to a level below that present at the start of the experiment prior to the addition of belinostat (Fig. 2a). DU145 and LNCaP cells were also sensitive to belionstat-mediated cytotoxicity following exposure for 3 days (data not shown). In contrast, belinostat at all concentrations examined seemed to be predominantly cytostatic on PC-3 cells during 3 days of exposure (Fig. 2b). Photographs taken of cells following exposure to belinostat were consistent with the cell counting results (Figs. 2a and b).
Effect of exposure time on belinostat growth-inhibitory and cytotoxic activity
Since even relatively high concentrations of belinostat seemed to be predominantly cytostatic on PC-3 cells during 3 days of exposure, we examined the effect of extending the exposure time beyond 3 days (Fig. 2c). The results of this analysis showed that belinostat was cytotoxic to PC-3 cells following drug exposure for 4–5 days. Thus, PC-3 cells require a somewhat longer exposure time than 22Rv1 cells (and DU145, LNCaP; data not shown) to undergo belinostat-mediated cytotoxicity, but are equally sensitive as other prostate cancer cell lines examined to belinostat-mediated growth-inhibition.
To further understand the effect of exposure time on belinostat activity, washout experiments were performed whereby belinostat (10 μM) was added to PC-3 cells and then removed at various times (Fig. 3a). Cell counts were then measured daily for 3 days to determine the effect of transient drug exposure on cell growth/viability. The results of this experiment indicated that a single exposure to belinostat for a relatively short period of time (<12 hr) produced a suboptimal effect on cell growth and that exposure to belinostat for 24 hr or greater resulted in nearly complete growth-inhibition.
We also performed experiments to determine whether cells exposed to belinostat for 48 hr were irreversibly growth-inhibited. To this end, PC-3 cells were exposed to belinostat (1 μM; Fig. 3b; 4 μM Fig. 3c) for 48 hr, washed and then counted daily for 3 days. The results of this experiment showed that PC-3 cells exposed to belinostat (1 μM) for 48 hr were able to recover and begin regrowing, whereas cells exposed to a higher concentration of belinostat (4 μM) did not begin to regrow within the timeframe analyzed.
Growth-inhibitory and cytotoxic activity of belinostat on prostate cancer cells and normal prostate epithelial cells
The clinical activity of an anticancer drug may be influenced by its therapeutic index which is a function of the differential effect of the drug on cancer vs. normal cells. To evaluate the activity of belinostat on prostate cancer cell lines vs. normal prostate cells, PC-3 cells and normal primary prostate epithelial cells (PREC) were examined using the trypan blue cell viability assay following exposure to belinostat (1, 4 μM) for up to 5 days (Fig. 4). Consistent with previous results (see above), belinostat completely inhibited the growth of PC-3 cells when viable cells were counted following 3 days of drug exposure, and there was evidence of cytotoxicity to these cells following 5 days of exposure to belinostat (4 μM) (Fig. 4a). In contrast, the effect of belinostat on PREC appeared predominantly cytostatic, even following 5 days exposure to 4 μM drug (Fig. 4b), a conclusion supported by subsequent cell-cycle analysis (see Fig. 5). In this experiment, PC-3 and PREC were cultured in different media. However, since similar results were obtained when these 2 cell types were cultured in the same media (data not shown) the use of different media does not explain the relative resistance of normal cells to belinostat-mediated cytotoxicity in this experiment.
Cell cycle effects of belinostat on prostate cancer cells and normal prostate epithelial cells
We next performed experiments to examine potential belinostat-mediated cell cycle effects on prostate cancer cell lines and normal prostate cells. Three prostate cancer cell lines (PC-3, DU145, LNCaP) and normal prostate PREC cells were exposed to belinostat for 24 or 48 hr and then stained with propidium iodide and analyzed by flow cytometry (Fig. 5).
On cancer cell lines, belinostat generally caused an increase in the percentage of cells in G2/M and a decrease in the percentage of cells in G0/G1 and S-phase, indicating that this compound induced growth arrest at the G2M phase of the cell cycle. For example, PC-3 cells showed an increase in the percentage of cells in G2/M (from 19% to 55%) and a decrease in the percentage of cells in S-phase (from 24% to 0%) and G0/G1 (from 41% to 30%) when comparing cells from drug-free control cultures to cells treated with belinostat (2 μM for 24 hr). Belinostat also increased the percentage of cells with subG1 DNA content (representing dead/dying cells), particularly in DU145 and LNCaP cells. For example, DU145 cells exposed to belinostat (2 μM for 24 hr) showed an increase in the percentage of subG1 cells from 3% (drug-free control) to 57% (belinostat-treated). PC-3 cells treated with belinostat (2 μM) for 24 hr did not show an increase in the percentage of cells with subG1 content relative to drug-free control cells, consistent with previous viable cell counting results (see Fig. 2c) which showed an absence of belinostat-mediated cytotoxicity on this cells line following only 24 hr of drug exposure.
Belinostat-mediated cell cycle effects on normal prostate PREC cells were qualitatively similar in some ways to those produced on prostate cancer cell lines, but generally less pronounced. For example, as was the case with cancer cell lines exposed lines to belinostat, PREC cells showed a decrease in the percentage of cells in S-phase (from 13% to 1%) and an increase in G2/M (from 12% to 26%) when comparing cells from drug-free control cultures to cells treated with belinostat (2 μM for 24 hr). However, in contrast to what was seen with DU145 and LNCaP prostate cancer cell lines, PREC cells showed almost no belinostat-mediated decrease in the percentage of cells in G0/G1 (from 70% to 66%), or increase in the percentage of cells with subG1 DNA content (from 5% to 7%) at this drug concentration/exposure time.
Activity of belinostat in an orthotopic prostate cancer xenograft model
The activity of belinostat in an in vivo setting was evaluated in an orthotopic prostate cancer model in which PC-3 cells were implanted into the prostates of nude mice. Following a treatment period of 3 weeks during which belinostat was administered i.p. twice a day (bid) at 20 or 40 mg/kg/dose, or 3 times a day (tid) at 40 mg/kg/dose, animals were sacrificed and tumors weighed (Fig. 6a). The results of this experiment demonstrated that belinostat effectively inhibited tumor growth in this model, with the 20 mg/kg (bid), 40 mg/kg (bid) and 40 mg/kg (tid) groups exhibiting tumor growth inhibitions of 27, 35 and 43%, respectively, relative to vehicle-treated animals (p < 0.001 for comparison of each belinostat-treated group to vehicle-treated group). Moreover, the 40 mg/kg (bid) group proved to be statistically different from the 20 mg/kg (bid) group (p = 0.028) and the 40 mg/kg (tid) group was statistically different from both the 20 mg/kg (bid) group (p < 0.001) and the 40 mg/kg (bid) group (p = 0.025). Weight loss did not exceed 10% in any group (data not shown).
In addition to the measurement of prostate tumor weights, the lungs of the animals in this experiment were macroscopically examined for evidence of gross metastasis. Interestingly, approximately half of the vehicle-treated animals possessed overt lung metastases, whereas gross metastases were not observed in any of the animals administered belinostat (Fig. 6b).
Antimigratory activity of belinostat on prostate cancer cells in vitro
Metastasis is considered one of the hallmarks of cancer8 and the metastatic process is believed to involve mechanisms such as the migration and invasion of tumor cells. Since belinostat appeared to have antimetastatic activity in the PC-3 orthotopic prostate xenograft model, we evaluated belinostat for potential antimigratory activity on the PC-3 tumor cell line in an in vitro assay. In this experiment, cells were treated with belinostat for 24 hr and then examined for migration across Transwell filters in response to 1% FBS (Fig. 7). The results of this experiment indicated that belinostat inhibited the migration of PC-3 cells in a dose-dependent fashion.
Molecular effects of belinostat on prostate cancer cells in vitro
Since tissue inhibitor of metalloproteinase-1 (TIMP-1) is a protein with demonstrated antiinvasive/metastatic properties that has been found to be downregulated in some prostate cancers,9–12 and since belinostat appeared to have antimetastatic activity in the PC-3 orthotopic prostate xenograft model, we examined the expression of TIMP-1 by immunoblotting of lysates from PC-3 cells exposed to belinostat and found this protein to be increased by belinostat (Fig. 8a; left panel, compare lanes 1 and 2). The belinostat-mediated increase in TIMP-1 expression was inhibited with Actinomycin D (Fig. 8a; left panel, lane 3) and thus required new RNA synthesis. Likewise, the belinostat-mediated increase in TIMP-1 expression was inhibited with cyclohexamide and emetine (Fig. 8a; left panel, lanes 4 and 5, respectively), and thus required new protein synthesis.
Since belinostat is an HDACi whose induction of TIMP-1 required new RNA and protein synthesis, we hypothesized that it might be possible to replicate the increase in TIMP-1 expression that was observed following belinostat exposure by employing siRNAs to various HDACs, focusing our investigation on class I HDACs which are generally ubiquitously expressed and detectable in the nucleus.1 To this end, PC-3 cells were treated with siRNA to HDAC1, 2 or 3 and evaluated for TIMP-1 expression via immunoblotting (Fig. 8a; right panel). This experiment showed that siRNA to HDAC3 effectively inhibited HDAC3 expression and strongly increased TIMP-1 expression. In contrast, siRNA to HDAC1 and 2 increased TIMP-1 expression only weakly, if at all (Fig. 8a), despite effective target knockdown in cells treated with these siRNAs (data not shown).
The p53 protein is frequently mutated in human cancers, including those of the prostate, and such mutations may impart gain-of-function properties on this protein.13, 14 We therefore performed an experiment to examine the effect of belinostat on the expression of p53 in the DU145 prostate cancer cell line, which is known to harbor mutant/activated forms of p53 that may provide a survival advantage to these cells.15 Lysates from DU145 cells exposed to belinostat for 48 hr were immunoblotted with an antibody to p53 (Fig. 8b; left panel) and the results of this experiment demonstrated that exposure to belinostat decreased the expression of mutant p53. We were unable to detect p53 expression in PC-3 cells (data not shown), consistent with literature reports demonstrating that this cell line harbors a p53 frameshift mutation and lacks detectable p53 protein expression.
Another oncogenic alteration that is becoming increasingly recognized as an important factor in prostate cancer is the overexpression of the ERG transcription factor due to gene rearrangement.16–18 ERG is one of a number of potentially oncogenic transcription factors belonging to the ETS family, and ∼40–50% of prostate cancers possess an ERG gene fusion. Moreover, the ERG rearrangement appears to be associated with a more aggressive form of prostate cancer and has been proposed as a prognostic factor for disease relapse.16, 19–21 Since the VCAP prostate cancer cell line has been reported to possess an ERG rearrangement fusion and elevated levels of ERG mRNA,18 we examined the effect of belinostat on ERG expression in this cell line and found that belinostat, even when used at a relatively low concentration (e.g., 0.33 μM) for just 24 hr, strongly decreased the expression of this protein (Fig. 8b; right panel). ERG expression was not observed in other prostate cell lines examined (data not shown), consistent with the absence of the ERG fusion or ERG overexpression in those cell lines.18 In contrast to its effects on ERG, belinostat increased the expression of the cell cycle inhibitory protein p21 in VCAP cells (Fig. 8b; right panel). The induction of p21 by HDACi, including belinostat3 has previously been described, and was included here mainly as a control to demonstrate that while some proteins (e.g., ERG) are decreased following exposure of VCAP cells to belinostat, others (e.g., p21) may be simultaneously increased.
In the present preclinical investigation, we found that belinostat potently inhibited the growth of prostate cancer cell lines in vitro. Cell cycle analysis showed that belinostat induced a large increase in the proportion of cells in G2/M, consistent with drug-induced growth-arrest in this phase of the cell cycle. In addition, belinostat was also cytotoxic to prostate cancer cell lines in vitro, as evidenced by a reduction in the number of viable cells to below starting cell numbers following exposure to belinostat and cell cycle data showing that belinostat exposure increased the proportion of cells with subG1 DNA content. Microscopic examination of cells following belinostat exposure also clearly indicated the cytotoxic nature of this compound on prostate cancer cell lines. Most prostate cancer cell lines showed significant cell death following exposure to belinostat for 1–3 days. Of the various prostate cancer cell lines analyzed in this study, PC-3 cells were the least sensitive to belinostat-mediated cytotoxicity, although even this cell line was susceptible to belinostat-mediated cytotoxicity following exposure for 4–5 days. Responsiveness to belinostat does not appear to be related to androgen sensitivity under our assay conditions since cell lines that are considered to be androgen-sensitive (LNCaP) exhibited similar responsiveness to belinostat as did cell lines that are considered to be androgen-insensitive (PC-3, DU145, 22Rv1). Normal prostate epithelial cells appeared to be generally less susceptible to belinostat-mediated growth-inhibition and cell death than prostate cancer cell lines, although clearly not entirely resistant to the effects of this drug.
Washout experiments indicated that exposure of prostate cancer cells to belinostat (10 μM) for short periods of time (e.g., <12 hr) was not as effective at inhibiting tumor cell growth as was exposure for longer periods of time (e.g., 24 hr). We have also found similar results on cell lines derived from cancer types other than prostate and by using cell viability assays other than trypan blue staining (data not shown). Exposure of prostate cells to belinostat for 2 days followed by drug washout indicated that tumor cells retained the capacity for regrowth following drug withdrawal in a manner dependent on drug concentration (cells exposed to 1 μM began regrowing while cells exposed to 4 μM did not). These findings support a scenario whereby belinostat interacts in a reversible manner with its target(s) and exposure of cells to a sufficient concentration of this drug for an appropriate period of time leads to irreversible growth-arrest and/or cytotoxicity. Consistent with the interpretation that belinostat interacts in a reversible manner with its target(s), this compound was previously shown to induce reversible histone hyperacetylation.3 Other HDACi of the hydroxamic class have also been shown to inhibit HDAC activity in a reversible manner.22 These findings indicate that maximal anticancer activity of belinostat may be achieved by maintaining a relatively high concentration of drug for a sustained period of time.
These in vitro findings, together with data obtained from a biodistribution study of radiolabeled belinostat in rodents which indicated the potential preference of this drug to localize to the prostate gland compared to most other tissues (data not shown), prompted us to examine the activity of belinostat in a prostate cancer orthotopic xenograft model. We chose to use PC-3 cells for this model even though this cell line is somewhat less sensitive to belinostat-mediated cytotoxicity than other prostate cancer cell lines since the conditions for the orthotopic model had been previously established using this cell line. In previous investigations, belinostat was administered once/day in animal efficacy models.3–6 However, the in vitro exposure/washout data obtained on this compound prompted us to investigate more frequent dosing schedules (bid, tid) in the present investigation with the goal of maintaining biologically active levels of belinostat over an extended period of time. A tolerability study (data not shown) in nude mice guided the selection of the doses of belinostat (20 and 40 mg/kg/dose) used in the present investigation. The maximal daily dose of belinostat administered to mice was 120 mg/kg (equivalent to ∼360 mg/m2) whereas the dose of belinostat that has been determined to be generally safe in humans and is being used in Phase II clinical trials is 1000 mg/m2. Thus, comparably more belinostat is used in humans than was administered to tumor-bearing animals in the present investigation. However, it should be pointed out that belinostat is generally administered once/day intravenously for 5 consecutive days in a 3 week cycle in the clinic and that direct comparisons between studies in mice and humans should not be made.
The finding in the present investigation that belinostat inhibited the growth of PC-3 orthotopic xenografts in a dose and schedule-dependent manner is consistent with the in vitro data which highlighted the importance of drug concentration and exposure time on belinostat-mediated growth-inhibition and cytotoxicity. An additional experiment on ovarian cancer xenografts comparing belinostat dosed at 60 mg/kg/dose once/day to this drug dosed at 20 mg/kg/dose 3 times/day confirmed the beneficial antitumor effect of dosing belinostat multiple times/day (data not shown).
In addition to reducing PC-3 tumor growth in the orthotopic model, none of the belinostat-treated animals developed gross lung metastases, whereas metastases were detected in nearly half of the vehicle-treated animals. Since only gross metastatic foci were enumerated in the present investigation, the potential presence of micrometastases in belinostat-treated animals cannot be ruled out. Also, the possibility that the absence of gross metastases in belinostat-treated animals is a function of the tumor growth-inhibition exhibited by this drug must be considered. However, it is also possible that the absence of gross metastases in belinostat-treated animals is due to a direct inhibitory effect of belinostat on the metastatic process. In support of the latter possibility, we found that belinostat inhibited the migration of PC-3 cells and increased the expression of TIMP-1 in these cells.
Tumor cell migration is an important step in the metastatic process and thus the ability of belinostat to inhibit migration may be functionally significant. It is unlikely that the ability of belinostat to inhibit the in vitro migration of PC-3 cells is due to potential cytotoxic effects of this drug on these cells since an exposure time of only 24 hr was used. Moreover, we have found that some class-selective HDACi inhibited the growth, but not the migration, of PC-3 cells (data not shown), thus demonstrating that the inhibition of cell growth and migration are separable and raising the possibility that the inhibition of a specific HDAC(s) may block cell migration. HDAC6 has been implicated in cell motility23, 24 and our findings that belinostat potently inhibits the enzymatic activity of this HDAC (IC50 of 82 nM on recombinant HDAC6; data not shown) and that some class-selective HDACi which reportedly do not effectively inhibit HDAC6 are ineffective inhibitors of PC-3 migration, support a scenario whereby belinostat-mediated inhibition of HDAC6 blocks PC-3 migration. We have also found that the belinostat-mediated inhibition of PC-3 migration is ameliorated by a compound called ITSA125 (data not shown), a compound that has been shown to suppress cell-cycle arrest mediated by the pan-HDACi TSA (but not by selective HDACi compounds devoid of HDAC6-inhibitory) and to inhibit TSA-mediated tubulin and histone acetylation. This finding is consistent with belinostat being a hydroxamate acid-based pan-HDACi and demonstrates that in addition to inhibiting the cell-cycle effects of pan-HDACi, ITSA1 also blocks the antimigratory activity of these compounds.
In addition to inhibiting migration, we have also shown that belinostat increases the expression of TIMP-1 in PC-3 cells in a manner requiring new RNA and protein synthesis. The observation that belinostat-mediated TIMP-1 induction is dependent on new RNA and protein synthesis is consistent with the known involvement of HDACs in gene regulation.1, 2 This finding, coupled with our results showing that TIMP-1 is strongly induced following siRNA-mediated knockdown of HDAC3, supports a scenario whereby HDAC3 participates in suppressing TIMP-1 transcriptional expression in PC-3 cells, and that this suppression can be relieved upon inhibition of this HDAC either pharmacologically with belinostat or genetically with HDAC3-specific siRNA. We cannot rule out the potential involvement of other HDACs in suppressing TIMP-1 expression, although knockdown of HDAC1 and HDAC2 caused only a weak, if any, induction of TIMP-1. Our finding that HDACi with selectivity for Class I HDACs also strongly induce TIMP-1 expression in PC-3 cells (data not shown) is consistent with our siRNA data implicating HDAC3 in regulating TIMP-1 expression.
Tumor cell invasion is believed to represent a requisite step in the metastatic processs,8 and TIMP-1 has been shown to block invasion by inhibiting the proteases that promote this activity.11, 12 There is support in the literature that an increase in the protease/inhibitor ratio compared to normal tissue exists in prostate cancer due to upregulation of proteases and downregulation of inhibitors9, 10 and thus the belinostat-mediated increase in TIMP-1 expression may alter the balance of proteases/inhibitors in favor of the inhibitors thereby leading to a reduction or elimination of invasion. We failed to detect TIMP family members other than TIMP-1 in PC-3 cells, and we have not thoroughly examined potential effects of belinostat on the expression of metalloproteinases in these cells (data not shown). Although most available evidence supports an antimetastatic role for TIMP-1 in prostate cancer, TIMP-1 has also recently been implicated as a survival factor in some cancers.26
Another finding of the present investigation is that belinostat caused a decrease in the expression of p53 in DU145 cells. Since these cells harbor gain-of-function mutated forms of p53 which may enhance tumor cell survival, the belinostat-mediated reduction of p53 in these cells may be functionally important. Mutant p53 has been shown to be a client protein of Hsp9027 and HDACi have been shown to disrupt Hsp90 function via inhibition of HDAC6.28 Since, belinostat potently inhibits the enzymatic activity of HDAC6 (see above), it is possible that the belinostat-mediated reduction of mutant p53 expression is caused via a reduction in Hsp90 chaperone function as a consequence of drug-induced HDAC6 inhibition. Consistent with the involvement of HDAC6 in this process, we have found that selective HDACi compounds devoid of HDAC6-inhibitory activity do not readily decrease the expression of mutant p53 in DU145 cells (data not shown). However, we have not generated any experimental data supporting the involvement of Hsp90 in this process.
In addition to mutant p53, another potentially oncogenic protein whose expression is decreased by belinostat in prostate cancer cells is the ERG transcription factor. As was previously discussed, ERG gene rearrangements leading to overexpression are common in prostate cancer and appear to be associated with aggressive disease. Interestingly, high HDAC1 expression has recently been associated with ERG-overexpressing prostate cancers,29 leading the authors of that study to suggest that ERG-positive prostate cancers may benefit from epigenetic therapy with HDACi. Moreover, ERG has been shown to indirectly interact with HDAC1.30 We now show that belinostat decreases ERG protein expression in a prostate cancer cell line that harbors an ERG gene fusion. This finding is consistent with the previously established connection between HDACs and ERG and with the proposal that HDACi may be particularly effective on ERG-positive tumors. Concurrent with the belinostat-mediated downregulation of ERG protein in VCAP cells, we also detected a simultaneous increase in p21 expression as a consequence of drug treatment. The p21 protein has been implicated in cell cycle inhibition and has previously been shown to be induced by belinostat as well as other HDACi.1–3
The relative contributions, if any, that the decrease in the expression of potentially oncogenic proteins such as ERG and mutant p53, and the increase in the expression of potentially growth-inhibitory proteins such as p21 make toward the overall growth-inhibitory/cytotoxic activity exhibited by belinostat on a particular cancer cell line is unknown. While it may be that the growth-inhibitory/cytotoxic activity of belinostat on most cancer cells is due to an accumulation of molecular alterations such as those described above, the possibility that belinostat induces a single dominant catastrophic molecular alteration which is primarily responsible for its activity across a wide variety of cancer cell lines, cannot be ruled out. In the latter scenario, many identified belinostat-mediated molecular alterations may in fact be superfluous and contribute little towards the growth-inhibitory or cytotoxic activity that this compound exhibits on most cancer cells. However, even if this scenario is correct, the fact that belinostat influences multiple pathways that may independently inhibit cancer cell growth may be beneficial as a way of providing “back-up” mechanisms to insure the death of cancer cells. Since belinostat and other HDACi compounds affect multiple independent pathways involved in the growth and/or survival of cancer cells, it may be more difficult for tumor cells to exhibit innate resistance or to acquire resistance to these agents compared to compounds which target a single pathway.
In summary, we have demonstrated that prostate cancer cell lines are susceptible to belinostat- mediated growth-inhibition and cytotoxicity, and that drug concentration, exposure time and cell line differences all influence belinostat activity. This compound has also been shown to reduce tumor growth and metastasis in an animal model of orthotopic prostate cancer, to inhibit tumor cell migration, and to decrease the expression of potentially oncogenic proteins that are associated with prostate cancer. Other HDACi have also shown activity in preclinical prostate cancer models.31–38 Together, these preclincial findings support the clinical evaluation of belinostat for the treatment of prostate cancer.